DSXchange

We would be happy, on a commercial basis, to analyze what your system requirements are and to recommend naming and other conventions based upon those findings. There is no "one size fits all" solution.

You could wrap that SQL in another SELECT statement to form the MIN and MAX of XX_MA_DT.

You could use an Aggregator stage to form the MIN and MAX.

You could use stage variables in a Transformer stage.

There's three ways, just off the top of my head.

The nice thing about standards is that there are so many to choose from.

For example, UNIX began life as a standard. So did SQL.

Searching the forum for "heap allocation failed" (exact match) could prove illuminating.

What has changed? ("Nothing" is not the correct answer. Clearly.)

Who has been fiddling with the project either at operating system level or TCL level? Find out what they did, and remedy it.

Can you specify MIN(TheDate) and MAX(TheDate) in your extraction SQL?

The simple answer is No. Get an ODBC driver for Lotus Notes. Even then it's not the easiest thing in the world to configure. An alternative strategy is to have some external process dump whatever it is you need from Lotus Notes into a text file format of some kind, and process that with DataStage.

Good catch - I was thinking transactional, and neglected consistency save points.

If the data are sorted on the grouping keys, and the Aggregator stage is informed of this fact, then the volumes are irrelevant - the stage can output a row (and free the memory associated with it) each time a sorted grouping key changes value - the sorting means that that value will never be seen a...

It's an Oracle restriction - nothing to do with DataStage. You have to specify (in DataStage) how to manage the indexes - you have two choices, essentially deferred update or drop and re-build.

This is well documented in the stage chapter in the Parallel Job Developer's Guide

All those comparisons can readily be performed in a parallel Transformer stage. Use stage variables to avoid unnecessary repetition of calculations and comparisons.

Read some more in Chapter 2 and 3 about data types. VarChar is stored with a prefix (an n-byte integer containing the actual length of the string part). You can see this in record schemas, for example string[max=255] {prefix=2} This indicates that a two-byte integer (actually a uint16) is ...

If you want to create a dummy file, then XML for Dummies is probably the best reference to consult!

Until and unless "they" find another mechanism than DSD.RUN to start parallel jobs, I can not agree.

Parallel Job Developer's Guide (version 7.5) page 3-33

It took less than a minute to search.